A new microbial polysaccharide membrane for ethanol dehydration by pervaporation

A new microbial polysaccharide membrane for ethanol dehydration by pervaporation

Journal of Membrane Science 425-426 (2013) 227–234 Contents lists available at SciVerse ScienceDirect Journal of Membrane Science journal homepage: ...

466KB Sizes 1 Downloads 407 Views

Journal of Membrane Science 425-426 (2013) 227–234

Contents lists available at SciVerse ScienceDirect

Journal of Membrane Science journal homepage: www.elsevier.com/locate/memsci

A new microbial polysaccharide membrane for ethanol dehydration by pervaporation Inˆes T. Meireles, Carla Brazinha, Joa~ o G. Crespo, Isabel M. Coelhoso n Requimte/CQFB, Departamento de Quı´mica, Faculdade de Ciˆencias e Tecnologia, Universidade Nova de Lisboa, Campus de Caparica, 2829-516 Caparica, Portugal

a r t i c l e i n f o


Article history: Received 6 June 2012 Received in revised form 31 August 2012 Accepted 1 September 2012 Available online 14 September 2012

In this study, a new microbial exopolysaccharide (EPS) membrane obtained from a low cost and abundant carbon source, the glycerol by-product of the biodiesel industry, was used for ethanol dehydration by pervaporation. Two types of membranes were prepared: homogeneous membranes obtained directly from the fermentation broth and crosslinked with tricloroacetic acid (EPS), and composite membranes with a thin film of crosslinked EPS supported in polyethersulfone (EPS-PES). These membranes were characterized in terms of their morphological structure, resistance to solvents and mechanical properties. They show a high affinity towards water, good chemical resistance towards organic solvents and adequate mechanical properties. Pervaporation experiments were performed using both types of membranes for different water concentration in the feed stream (5.0 to 10.0 wt%) at a constant temperature of 30 1C. The membranes developed, in particularly the composite EPS-PES membrane, exhibit a high potential for ethanol dehydration, since a water/ethanol selectivity of 3000 at 5.0 wt% water concentration in the feed was achieved. Considering the good results obtained, these new membranes will be also evaluated for the dehydration of other industrially relevant solvents, for separation of polar and non-polar compounds in organic mixtures, and also for solvent-resistant nanofiltration processes. & 2012 Elsevier B.V. All rights reserved.

Keywords: Microbial polysaccharides Biopolymers Ethanol dehydration Hydrophilic pervaporation Solvent dehydration

1. Introduction Solvent dehydration processes have a high economic and environmental relevance in the pharmaceutical, fine-chemistry and chemical industry. Among the available techniques for solvent dehydration, hydrophilic pervaporation has gained a progressive recognition due to its relative simplicity and relative low energy input, as compared to traditional energy-intensive methods [1]. Additionally, pervaporation is able to resolve solvent– water azeotropic mixtures, which cannot be easily achieved by traditional distillation processes, and require the use of entrainers, leading to higher costs and complexity. Particularly, the dehydration of alcohols by hydrophilic pervaporation has been studied intensively, representing by far the main industrial application of pervaporation [2–4]. The process of hydrophilic pervaporation involves three sequential steps: selective sorption of water from the feed solution to the membrane, diffusion of water/solvent through the membrane and desorption at the permeate side [5]. Transport through a non-porous membrane is usually explained by the


Corresponding author. Tel.: þ351 21 2948302; fax: þ351 21 2948385. E-mail address: [email protected] (I.M. Coelhoso).

0376-7388/$ - see front matter & 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.memsci.2012.09.002

sorption equilibrium of the feed components and by the mobility of the permeating compounds across the membrane. Therefore, if the selective transport of a target solute from the feed stream is aimed, it is necessary to assure that the membrane presents a good compromise between flux and selectivity for the target solute, as well as high mechanical and chemical stability for the target application [6–8]. Hydrophilic polymers, such as poly(vinyl alcohol) (PVA), poly(acrylic acid) and polyacrylonitrile have been selected, among others, as membrane materials for the dehydration of various solvents [3,9–11]. Nowadays, with the growing environmental awareness, it is becoming progressively relevant the development of membranes from renewable resources, namely using biopolymers produced from agro-industrial by-products. This approach will present several environmental and economic advantages in the longterm, reducing the impact of low-value agro-industrial wastes and transforming them into valuable materials. On the other hand, nonrenewable resources, including petroleum resources, are becoming more expensive, therefore alternatives are needed for polymer production [12,13]. Several authors have proposed the use of biopolymers for solvent dehydration by pervaporation using polysaccharide based membranes, such as chitosan and sodium alginate [14–16]. These polymers are commonly extracted from natural resources and


I.T. Meireles et al. / Journal of Membrane Science 425-426 (2013) 227–234

the availability and quality of the biopolymer sources depend from climate conditions and season of the year, which leads to variability of the polymer quality. Microbial polysaccharides represent a better alternative, since the production parameters can be easily controlled and high production rates can be obtained. However the high cost of most carbon sources studied, in particularly sugars, has a high impact on the production costs, limiting the potential use of those microbial polymers [17]. Therefore, attention has been directed to the search for viable low-cost carbon sources, namely industrial and agro-industrial by-products. Recently, it has been reported that by-product streams containing glycerol, from biodiesel production, could be directly applied as carbon source for the production of a new extracellular polysaccharide (EPS), using Pseudomonas oleovorans, with attractive productivity and polymer yield. This new EPS is a high molecular weight (4.6  106 g/mol), negatively charged heteropolysaccharide, composed by sugars (68% galactose, 13% glucose, 17% mannose and 2% rhamnose) and acyl groups (3%), such as pyruvil, succinyl and acetyl [18]. It was also demonstrated that this polymer can be used for producing films which are insoluble in most common organic solvents (actually they revealed to be insoluble in all solvents tested, namely ethanol, isopropanol, dichloromethane, acetone, ethyl acetate and tetrahydrofuran). Moreover, when cross-linked with appropriate agents, these polysaccharide films become insoluble in aqueous solutions while maintaining their resistance to organic solvents. In addition to their structural advantages, these membranes have an important economic advantage because they can be produced by bacteria that use glycerol, an abundant and low cost carbon source [19]. This work is focused on the development and characterization of homogeneous membranes of EPS obtained directly from the fermentation broth and, also, of composite membranes prepared by casting a thin-film of EPS in a porous support of polyethersulfone (PES). The EPS-PES composite membranes were developed in order to provide a higher flux of the target solute–water, while keeping the selectivity of the separation process. Additionally, it is aimed to obtain membranes with adequate mechanical properties for use under industrial operating conditions. The membranes developed were characterized in terms of their structure, by optical and scanning electron microscopy (SEM), and their interaction/resistance to different organic compounds and water, by determining their swelling degree. The mechanical properties of the different materials involved – homogeneous EPS membrane, polyethersulfone support, and composite membrane of PES-EPS – were also evaluated. Finally, pervaporation experiments were performed using ethanol solutions with different water contents (from 5.0 to 10.0 wt%), typically in the concentration range of interest for ethanol dehydration. The results obtained are interpreted and compared with other biopolymer membranes and with commercial membranes.

2. Experimental 2.1. Materials The PES membrane, used as support of the composite membrane, was obtained from Gelman Sciences SB-6407 (USA). This membrane presents a nominal pore size of 0.45 mm, positive charge (1.15 meq/cm2) and area of 17.35 cm2. The commercial PERVAPs 4101 membrane, commonly used for ethanol dehydration, was provided by Sulzer Chemtech (Switzerland). Isopropanol (99.9%), ethanol (99.8%) and ethyl acetate (99.9%) were purchased from Panreac Quı´mica SA (Spain). Dichloromethane (99.9%) was ¨ (Germany) and Sigmaobtained from Fluka Riedel-de Haen Aldrich Chemie GmbH (Germany), respectively. Tetrahydrofuran (THF) (99.8%) was obtained from Merck (Germany). Acetone (99.8%) was purchased from Valente e Ribeiro Lda. (Portugal). Molecular sieves 4 A˚ (8–12 mesh), used to dry the solvents prior to the swelling experiments, were purchased from Sigma-Aldrich Chemie GmbH (Germany). Trichloroacetic acid (TCA), used to crosslink the EPS membranes, was obtained from Merck (Germany).

2.2. Membrane preparation Two types of membranes were prepared: homogeneous membranes of exopolysaccharide (EPS) produced by fermentation of P. oleovorans, and composite membranes of EPS casted on a polyethersulfone (PES) support. The EPS homogeneous membranes were prepared directly from the fermentation broth diluted with deionised water (1:2 v/v), followed by cell separation by centrifugation (13,000 rpm, 30 min, 10 1C). An aqueous solution of trichloroacetic acid (TCA, 100% w/w) 1:10 (v/v) was added to the supernatant and, after remaining 10 min at 4 1C, a second centrifugation was performed to remove the precipitated proteins. The polymer extraction was accomplished using acetone 1:3 (v/v). The resultant precipitated biopolymer (0.49% dry weight [19]) was collected and immediately dissolved in an aqueous solution of sodium azide (10 ppm) to prevent microbial growth. The membrane was cast in a Teflon plate and dried at 30 1C during 48 h. The composite EPS-PES membranes were prepared by the deposition of four EPS layers on a PES support, following a procedure similar to the one used for preparing the EPS homogeneous membranes. In order to help the spreading of the layers, an increased polymer concentration was used. Particularly, the dilution of the fermentation broth with deionized water was of 1:1 (v/v), instead of 1:2 (v/v) as in the case of the homogeneous membrane. Each layer has an amount of approximately 5 g by total area of membrane prepared, and it was allowed to dry for 6 h at 30 1C, before adding the next layer.

Fig. 1. Pervaporation (PV) experimental set-up: 1 – closed stirred feed vessel; 2 – recirculation pump; 3 – flat circular pervaporation module; 4 – pressure transducer; 5 – temperature sensor; 6 – condenser with liquid nitrogen; and 7 – vacuum pump.

I.T. Meireles et al. / Journal of Membrane Science 425-426 (2013) 227–234

2.3. Membrane characterization 2.3.1. Morphological structure Scanning electron microscopy (SEM). The morphological structure of the membranes was acquired by scanning electron microscopy (SEM) at 10 kV with a Jeol JSM-7001F, Field Emission scanning electron microscope. Optical microscopy. The optical microscope (Axioplan 2ie imaging, Zeiss, Germany) was used in order to measure the thickness of the membranes (homogeneous and composite membrane of EPS, and composite PERVAPs 4101), without exposing them to vacuum conditions required by SEM technique. For the homogeneous membrane and for the composite membrane visible light in dark field an amplification of 50 was used. And, for the composite PERVAPs 4101 membrane UV light an amplification of 50 was used. 2.3.2. Solvent resistance and swelling behaviour Small square pieces of homogeneous membranes of EPS (2 cm  2 cm) were dried during 24 h at 70 1C. After this procedure the membranes were weighed with a digital balance (KERN ABJ, Germany). Then, the membranes were immersed in selected organic solvents (ethanol, ethyl acetate, isopropanol, dichloromethane, acetone, tetrahydrofuran (THF)) and water for 24 h at room temperature (24.671.1 1C). In order to ensure that the organic solvents were completely dried, molecular sieves were used to adsorb water from these solvents before each experiment. After this procedure, the membranes were taken out and the excess of solvent was removed from their surfaces using tissue paper and weighed. The degree of swelling for each membrane was determined by Eq. (1): Swelling w ð%Þ ¼

W f W i  100 Wi


where Wf, and Wi are the weights of the swollen and dry membranes, respectively. In all experiments, the membranes kept their integrity and no dissolution to the contacting solvents was apparent.


25.070.5 1C. After this period it could be considered that the membranes were in equilibrium with the contacting media. The membranes were then removed and weighed after the excess of solvent was removed with tissue paper. The increase in membrane weight is due to the water or ethanol sorbed by the membrane. In order to determine accurately the amount of water and ethanol sorbed, the membranes were heated in a control mode using a thermogravimetric balance (TGA-50) at 39.370.9 1C, and the mass loss was recorded along time. The sorption coefficients of water and ethanol were obtained in gi/g wet membrane. 2.5. Pervaporation experiments Pervaporation experiments were carried out for both types of membranes, using the experimental setup shown in Fig. 1. This experimental setup consists essentially of a closed feed vessel, a flat circular pervaporation module, a condenser and a vacuum pump (Edwards RV12, UK). The feed vessel was closed and operated with a small headspace. It has a relatively large volume (600 ml) to membrane area ratio, in order to minimize ethanol depletion from the feed during each experiment (which was experimentally confirmed to be lower than 0.2 wt%). A heating bath (Julabo EH,Germany) connected to the feed vessel was used to keep the temperature of the feed solution constant at 30.0 7 0.3 1C. The feed solution was pumped by a recirculation pump (Gilson Minipul S3, USA) to the pervaporation module (flat circular cell) at a flow rate of 22 ml/min. The surface area of the membrane in contact with the feed mixture was 7.54 cm2. The temperature of the pervaporation module was maintained constant by a heating belt (RKC CB100, Japan) at 30.0 7 0.1 1C. A cold trap immersed in liquid nitrogen was used to collect the permeated vapor. The vacuum at in the downstream circuit was 0.9 7 0.1 mbar obtained by the vacuum pump. Different concentrations of water in the feed solution were studied (5.0 wt% and 10.0 wt% for the homogeneous membrane, and 5.0 wt%, 7.5 wt% and 10.0 wt% for the composite membrane). Each experiment lasted 9 h for the homogeneous membrane. For the composite membrane each experiment lasted only 6 h, because the amount of permeate recovered was large enough to be analysed. 2.6. Analytical methods

2.3.3. Mechanical properties These studies were performed in order to characterize the potential behaviour of the membranes under study when exposed to the operating conditions in a pervaporation module. Mechanical properties such as tensile strength and elongation at break were measured using a TA-Xtplus texture analyser (Stable Micro Systems, Surrey, England). The tensile stress at break is defined as the force that is required to break the films and the elongation at break is calculated as the percentage increase in length that occurs before the film breaks under defined tension conditions. The membranes were placed between the grips and stretched at a rate of 0.5 mm/s in tension mode. All mechanical tests were performed at 22.072.0 1C in two different conditions, which aimed to mimic the environment of the membrane before and after being used in pervaporation experiments: membrane samples were equilibrated during 4 days at 44.3% relative humidity (common atmospheric conditions); they were also immersed during 24 h in a solution of 10.0 wt% water in ethanol. At least three samples of each membrane were used for each test.

2.6.1. Gas chromatography analysis The concentration of ethanol in permeate was determined by gas chromatography (GC, CP-3800, Varian, USA) using an auto sampler (Combi PAL, CTC Analytics, Switzerland). The samples were injected into the gas chromatograph using the technique of static headspace. Specifically, 1 g of sample was heated for 10 min at 70 1C in vials hermetically sealed, followed by the injection of 1 ml of headspace sample in the gas chromatograph. The injector temperature was 170 1C and the FID (flame ionization detector) temperature was 250 1C. The split ratio used was 10. The carrier gas was Helium and the standard deviation for the different compounds examined was found to be always lower than 5%. The column temperature was programmed as follows: 35 1C during 5 min followed by an increase of temperature of 10 1C/min until 215 1C [20].

2.4. Sorption experiments

3.1.1. Membrane structure The top layer surface and the cross-section views of the homogeneous membrane of EPS and the composite membrane of EPS-PES are shown in Fig. 2.

Membranes with a known mass were immersed in pure deionised water and also in pure ethanol (99.8%) for 24 h at

3. Results and discussion 3.1. Membrane characterization


I.T. Meireles et al. / Journal of Membrane Science 425-426 (2013) 227–234

Fig. 2. SEM views: (a) cross-section image of an EPS homogeneous membrane; and (b) cross-section image of a EPS-PES composite membrane.

The homogeneous membrane image shows a dense surface without defects. The cross-section image of the composite membrane shows the porous support of PES and the dense top layer of EPS (also without noticeable defects). It can also be noticed that there is a perfect separation and yet a good adhesion, between the layers of EPS and of the support of PES. In order to determine the thickness of the membranes, without the vacuum effect present during the analysis by SEM, they were also visualized by optical microscopy. Particularly, the thickness of the membranes was obtained through an optical microscopy image analysis using the AutoCADs software programme. It was found that the homogeneous membranes present a thickness of 17.973.4 mm, and the composite membranes present a total thickness of 164.5 72.8 mm, from which 12.570.4 mm correspond to the thickness of the active layer. The active layer of the PERVAPs 4101 was also measured by optical microscopy and a thickness of 3.570.1 mm was obtained.

3.1.2. Solvent resistance and swelling behaviour These experiments were performed with selected solvents, commonly used in diverse industrial applications, and water (see Table 1). As mentioned, in all experiments performed, the

membranes kept their integrity and no relevant dissolution to the contacting solvents was apparent. This behaviour of the films in water might be unexpected, since the films were produced through dissolution of the EPS in water. However, it was reported that the addition of TCA in the purification step lowered the pH of the solution, promoting the formation of strong interactions between polymer chains and making the membranes resistant to liquid water. Crosslinking reactions are taking place only during the drying process, because polymer molecules are driven to closer contact, as water evaporates [19]. From the increase of membrane weight, after contact with each solvent, it may be inferred if a given solvent presents a high affinity towards the membrane and solubilizes within its polymeric structure. From Table 1 it may be concluded that the homogenous membrane exhibits a negligible swelling when in contact with most solvents, isopropanol and dichloromethane being the organic solvents that promote a higher swelling. The negative swellings observed in the case of acetone and ethyl acetate were due to a minor leaching of the membrane. Considering the low value obtained for ethanol and the high swelling observed for water (swelling of 2.4 wt% and 95.9 wt%, respectively for ethanol and water), it may be concluded that this membrane exhibits a

I.T. Meireles et al. / Journal of Membrane Science 425-426 (2013) 227–234

Table 1 Swelling results of the homogeneous membrane of EPS at 23 7 1 1C.

 5.3  1.3 2.4 2.8 7.2 9.3 95.9

high relative affinity for water, which suggests a potentially good behaviour for the selective transport of water against ethanol. The low degree of swelling in the presence of most organic solvents also suggests that this membrane may be extremely stable when in contact with these solvents. Also, it a low permeation of these solvents through the membrane may be expected, making it a good candidate for solvent dehydration.

3.1.3. Mechanical properties Mechanical strength tests were carried out for the two types of membranes, homogeneous EPS and composite membrane of EPSPES, and also for the porous PES supporting membrane. Two different experimental conditions were selected in order to: mimic the exposure of these materials to ambient conditions, by equilibration with air with a relative humidity of 44.3% at 22 1C; and mimic conditions similar to the ones occurring during ethanol dehydration, by immersion during 24 h in a solution of 10.0 wt% water in ethanol. Tensile stress studies were performed, instead of compressive stress tests, because the former reflect better the situation of a membrane during operation when assembled in a pervaporation module. Fig. 3 shows a sample of stress–elongation curves obtained with these materials. Representing the stress as a function of elongation (%), it is possible to obtain the Young’s modulus by determining the slope in the linear region, as shown in Fig. 3. The stress at break relates with the ability of the material to resist rupture when a tension force is applied, and the elongation at break is the elongation recorded at the moment of rupture of the specimen, expressed as a percentage of the original length. Furthermore, the Young’s modulus describes tensile elasticity, or the tendency of an object to deform along an axis when opposing forces are applied along that axis. A high Young’s modulus refers to a material with a low tensile elasticity. The results obtained are summarised in Table 2. Analysing Table 2 and Fig. 3 it can be observed that the homogeneous EPS membrane exhibits, for both environmental conditions tested, a behaviour that reflects a higher resistance to rupture and a lower elasticity than the PES support material. As it can be seen, the EPS membrane presents a significantly higher stress at break (higher resistance to rupture), and a higher Young’s modulus and lower elongation at break, which reflect a lower elasticity. From the data obtained it can be also concluded that, when immersed in the 10.0 wt% water in ethanol solution, all materials tested exhibit a lower resistance to break and a higher elasticity. These results indicate that the membrane becomes more flexible due to solubilisation of the media components within the polymer matrix, leading to a rearrangement of the polymeric chains and, ultimately, swelling. Similar results were also observed by Zereshki et al. [21] with a composite membrane of poly(lactic acid)/poly(vinyl pyrrolidone).


Stress (MPa)

Acetone Ethyl acetate Ethanol THF Isopropanol Dichloromethane Water

EPS membrane y = 1167.9x + 1.8 R² = 0.999


44.3% RH 10 wt% water

10 5 0 0.0%

y = 683.9x - 0.3 R² = 0.999


2.0% Elongation (%)



PES support

4.5 4

y = 60.7x - 0.3 R² = 0.999

3.5 Stress (MPa)

Swelling (wt%)


3 2.5

44.3% HR


10 wt% water

1.5 1 0.5 0 0.0% 7

y = 38.1x + 0.2 R² = 0.999


20.0% 30.0% Elongation (%)



EPS-PES composite membrane

6 Stress (MPa)




5 4

y = 285.9x - 0.1 R² = 1.000

44.3% HR


10 wt% water

2 1 0 0.0%

y = 73.1x + 0.1 R² = 0.999





Elongation (%) Fig. 3. Representative stress–elongation curves and Young’s modulus (slope of linear equations inside boxes) of the various materials exposed to air with 44.3% relative humidity at 22 1C and immersed in 10.0 wt% water in ethanol solution: (a) EPS membrane; (b) PES support; and (c) EPS-PES composite membrane.

Table 2 Average values obtained for the mechanical properties of the membranes developed and for the PES supporting material, for the two experimental conditions studied: exposure to air with 44.3% relative humidity at 22 1C and immersion in 10.0 wt% water in ethanol solution. Membrane Conditions Stress at break Elongation at (MPa) break (%) EPS

44.3% RH water 10.0 wt%

28.7 72.6 19.7 70.0

2.4 7 0.3 2.8 7 1.0

PES support

44.3% RH water 10.0 wt%

3.5 70.4 3.4 70.1

18.9 7 4.4 35.1 7 10.2


44.3% RH water 10.0 wt%

6.8 71.6 2.8 70.6

3.5 7 1.3 14.7 7 4.7

Young’s modulus (MPa) 1275 7 150 991 7 430 597 3 527 20 254 7 20 607 20


I.T. Meireles et al. / Journal of Membrane Science 425-426 (2013) 227–234

For the composite membrane two different behaviours were observed, depending on the experimental conditions used. In the case of the membrane exposed to air with 44.3% of relative humidity, it can be observed that this membrane exhibits an intermediate behaviour between the homogeneous EPS membrane and the PES supporting material. The stress at break and the Young’s modulus present an intermediate behaviour between the homogeneous membrane of EPS and PES support, although, the elongation at break is similar to the EPS membrane. When the membrane is immersed in a solution of 10.0 wt% water in ethanol, the support (PES layer) seems to dominate the mechanical properties of the composite membrane. Since these environmental conditions are similar to the ones a membrane experiences during ethanol dehydration by pervaporation, it may be concluded that the composite membrane developed will present a mechanical behaviour close to a PES membrane, which assures an adequate mechanical performance during operation. The mechanical properties of EPS membranes developed in this work can be compared with the results for other dense membranes of biopolymers reported in the literature (Table 3). From Table 3 it appears that the EPS-based membranes present one of the highest Young’s modulus, which means that these membranes are more rigid. The values of stress at break for the EPS membrane are similar to the membrane of sodium alginate (28.7 MPa and 21.2 MPa, respectively), but lower than chitosan. These results indicate that the homogeneous membrane presents similar mechanical properties when comparing with other biopolymers referred in literature. It is difficult to establish a thorough comparison with other biopolymers because, in most cases, the mechanical properties were not measured under similar conditions. 3.1.4. Pervaporation experiments The pervaporation results obtained for the homogeneous and composite membranes are presented in Table 4. From Table 4 it can be concluded that both membranes present a high selectivity for water against ethanol. As expected, when the driving force for water transport increases (higher concentration of water in the feed stream), higher fluxes of water are obtained. The comparison between the homogeneous and the composite membrane is more difficult to establish because the top layer of the composite membrane was prepared with a solution with a higher polymer concentration (1.24 g/cm3, versus 0.72 g/cm3) leading to the formation of an active layer denser than the homogeneous membrane. For this reason, the fluxes obtained with the composite membrane were lower. In order to better quantify the performance of these membranes, the permeability and selectivity values must be calculated. Permeability of each feed component may be calculated by considering the solution-diffusion equation [23]: Ji ¼



ðpi,f eed pi,perm Þ3J i ¼



ðgi,f eed xi,f eed psat i,f eed pi,perm Þ

pressure (Pa), d is the thickness of the membrane or dense layer of the composite membrane (m), g is the activity coefficient of a target compound (dimensionless), x is its molar fraction in the feed liquid (dimensionless) and psat is the saturation vapor pressure (Pa). For the range of compositions studied the saturation vapor pressures at 30 1C [24], calculated through the Antoine’s equation, and the activity coefficients are presented in Table 5. The values of the activity coefficients were calculated by the Wilson method, as in [20]. This method allows for calculating values of activity coefficients in the whole range of compositions studied, taking into account the non-ideality of liquid mixtures due to specific interactions, introducing the concept of local composition [25]. This method was also selected because the values calculated are in agreement with experimental values obtained for water and ethanol at infinite dilution. The permeability values, calculated with Eq. (2), and the selectivity values (calculated as the ratio between the permeability of water and ethanol) allow for better understanding the intrinsic performance of the membranes under study. These parameters are used to compare the results obtained under different experimental conditions [26]. The results obtained are shown in Table 6. From Table 6(a) it can be observed that, for both membranes developed (homogeneous and composite), the permeability values increased and the selectivity values decreased, with increasing water concentration in the feed. These permeability values may be explained by a rearrangement of the membrane polymer, which becomes more flexible (in line with the decrease of the Young’s modulus, see Section 3.1.3.) as the concentration of water in the feed phase increases. A higher concentration of water in the feed leads to a higher concentration of water within the

Table 4 Mass fluxes of water and ethanol for the homogeneous membrane of EPS and for the composite membrane of EPS-PES. Membrane

[Water]feed (%wt)

Jt (g/m2 h)

Jw (g/m2 h)

Jet (g/m2 h)


5.2 10.1 5.2 7.6 10.4

21.7 7 0.2 59.1 7 0.4 10.9 7 0.0(2) 20.5 7 0.0(4) 39.3 7 0.2

18.2 70.1 50.5 70.2 10.8 70.0(1) 20.2 70.0(2) 36.0 70.1

3.57 0.1 8.67 0.2 0.17 0.0(1) 0.37 0.0(2) 3.37 0.1


Table 5 Values of calculated saturation vapor pressure at 30 1C (a) and activity coefficients (b) for water and ethanol. (a) Compound

psat (Pa) i

Water Ethanol

4276.2 10462.0

(b) [Water]feed (%wt)



5.2 7.6 10.4

1.04 1.05 1.07

2.33 2.15 1.98


where i is a target compound (water or ethanol), J is the molar flux (mol/m2 s), P is the permeability (mol/m s Pa), p is the partial

Table 3 Comparison of the mechanical properties of dense membranes composed by biopolymers. Membrane

T (1C)

Stress at break (MPa)

Elongation at break (%)

Young’s modulus (MPa)



22 22

28.7 7 2.6 6.8 71.6

2.47 0.3 3.57 1.3

1242 7 130 254 7 20

Present study at 44.3% RH



82.4 7 8.5

5.27 0.9

534 7 44

Suyatma et al. [22] at 50% RH

Sodium alginate Chitosan-Sodium Alginate

– –

21.2 72.3

6.9 6.3

– –

Kanti et al. [16] at 0% RH

I.T. Meireles et al. / Journal of Membrane Science 425-426 (2013) 227–234


Table 6 Comparison of permeability and selectivity values of water and ethanol for the homogeneous membrane of EPS and for the composite membrane of EPS-PES (a) and for other membranes referred in literature (b). (a) Membrane

[Water]feed (%wt)

Pw  1012 (mol/m s Pa)


5.2 10.1

4.37 0.8 7.67 1.5

4.07 0.8 10.77 2.1

108 69


5.2 7.6 10.4

1.87 0.3 2.67 0.4 3.87 0.6

0.0(6) 7 0.0(1) 0.2(5) 7 0.0(4) 2.97 0.5

3000 1040 134

(b) Membrane

[Water]feed (%wt)

T (1C)


5.2 5.0 10.4 10.0

30 70 30 45

PERVAPs 4101



EPS Active layer of EPS-PES

Spure water (gw/(gwet 0.443 0.481


Pet  1014 (mol/m s Pa)

Selectivity (w-et)


12.5 – 12.5 1.8

1.8 4.7 3.8 2.5

0.0(6) 0.2(4) 2.9 –

3000 1950 134 –





Present study Namboodiri and Vane [27] Present study Namboodiri and Vane [27] refers to [28] Present study

Spure ethanol (get/(gwet

Table 8 Ratio between water and ethanol permeabilities and ratio between water and ethanol sorption coefficients, calculated in a molar basis, for both membranes. Membrane

[Water]feed (%wt)


5.2 10.1

108 69


5.2 7.6 10.4

3000 1040 134




0.034 0.058

membrane and, therefore, to membrane swelling and higher mobility of the polymer chains. As a consequence, permeability increases and selectivity decreases. Still, it must be stressed the high values of selectivity for water transport against ethanol in range of concentrations relevant for ethanol dehydration. When comparing the composite membrane with other membranes reported in the literature (Table 6b) it can be observed that the EPS-PES membrane has an extremely high selectivity at 5.0 wt% water in the feed, which compares favourably with the results reported by Namboodiri and Vane (selectivity of 3000 and 1950 for EPS-PES and PAAHCl-PVA-PVDF, respectively). This comparison should be established with care because these works were not performed at the same temperature. A higher temperature (70 1C) will favour flux of the EPS-PES membrane but, probably, at the expenses of a decrease in selectivity. On the other hand, if we compare the composite membrane of EPS with a composite membrane also composed by biopolymers (the sodium alginate-PVA membrane), it can be observed that the EPS-PES membrane exhibits a higher permeability for water than the sodium alginate-PVA, at 10.0 wt% water in the feed. However, the commercial membrane of Sulzer (PERVAPs 4101) presents a higher value of water permeability and a lower value of ethanol permeability, at 10.0 wt% water in the feed. Considering these results it is important to identify the role of sorption and diffusion in the transport of ethanol and water, in order to direct further membrane development efforts in the right direction. Transport of water and ethanol through exopolysaccharide dense membranes. The permeability of a given compound i (Pi) may be expressed as the product between its sorption coefficient towards the membrane (Si) and its diffusion coefficient within the membrane (Di): P i ¼ Si  Di

Selectivity (w-et)

Pw  1012 (mol/m s Pa)

d active layer (lm)

Table 7 Sorption coefficients for water and for ethanol, for the homogeneous membrane of EPS and for the active layer of the composite EPS-PES membrane. Membrane

Pet  1014 (mol/m s Pa)




The sorption coefficients of pure liquid water and ethanol, Spure , were obtained for the homogeneous membrane and for the i active layer of the composite membrane by thermogravimetry. Sorption coefficients for water and ethanol using binary mixtures with the compositions of interest were not experimentally measured but, for the purpose of this discussion, such values are not vital. The values obtained are presented in Table 7. These results show clearly that water is preferentially sorbed by both membranes, as expected. The denser EPS layer of the EPS-PES composite membrane presents a slightly higher sorption coefficient for water (9% higher), but also a higher sorption coefficient for ethanol (70% higher), than the homogeneous membrane. Unlike sorption coefficients, diffusion coefficients of ethanol are expected to decrease more significantly than the diffusion coefficients of water in a denser top-layer, since diffusion is related with the size of the molecules. Indeed, since ethanol has a higher molecular mass than water, its diffusion coefficient is expected to be more affected in a more compact membrane. Actually, the permeability results obtained (see Table 6a) show exactly this behaviour, which explains the higher selectivity of the composite membrane for water transport in comparison with the homogeneous membrane. pure If we compare the ratio of sorption coefficients (Spure w /Set ) (calculated on a molar basis, from the values shown in Table 7 expressed on a mass basis) and the corresponding ratio of permeabilities (Pw/Pet), we may draw conclusions about the relative importance of each step (sorption and diffusion). The results obtained can be observed in Table 8. From this table it may be concluded that, for the homogeneous EPS membrane, transport is mainly controlled by the strong affinity of water to the membrane, while the contribution of diffusion to the overall selectivity seems less important. On the other hand, when analysing the values obtained for the composite EPS-PES membrane,


I.T. Meireles et al. / Journal of Membrane Science 425-426 (2013) 227–234

it seems clear that the transport process is dominated by diffusion phenomena for lower water content in the feed stream but, progressively, as the water content is increased in the feed stream, sorption becomes the controlling mechanism. For the composite membrane the decrease of the diffusion step control with increasing water concentration in the feed, corresponds to a higher plasticization of the polymeric structure, as discussed above, leading to a less hindered and selective diffusion. These results suggest that future developments should be focussed on identifying the optimal degree of cross-linking of the membrane, in order to assure high diffusion selectivity without compromising the absolute value of water permeability. Additionally, the thickness of the composite membrane should be also optimised in order to guarantee a good membrane permeance without affecting the good chemical and mechanical properties observed.

4. Conclusions This work proposes a new membrane material, obtained from a renewable source, for ethanol dehydration by pervaporation. This microbial exopolysaccharide is produced using a low cost, abundant carbon source, the glycerol by-product of the biodiesel industry, which represents a key economic advantage. The produced EPS homogeneous membrane, obtained directly from the fermentation broth and crosslinked with tricloroacetic acid, and the composite membranes of EPS with a porous polyethersulfone as support (EPS-PES) were characterized in terms of their morphological structure, solvents resistance and mechanical properties. The membranes showed a high affinity to water, good chemical resistance towards organic solvents and adequate mechanical properties under conditions similar to pervaporation. These new membranes may become an interesting alternative to commercial hydrophilic pervaporation membranes for the dehydration of ethanol. Particularly, the composite EPS-PES membrane, exhibited a water/ethanol selectivity of 3000 at 5.0 wt% of water in the feed. Strategies, for further improvement should include optimisation of polymer concentration and crosslinking conditions, in order to improve selectivity for higher concentration of water in the feed stream (10.0 wt% of water in the feed) without compromising the absolute value of water permeability. The preparation of these EPS membranes will also be optimised in order to render their fabrication more sustainable, namely by eliminating the use of acetone or any other hazardous organic solvent. Additionally, it will be important to evaluate these new membranes in other industrially relevant processes, such as the separation of polar and non-polar compounds in organic mixtures, and also for solvent-resistant nanofiltration processes.

Acknowledgements The authors are grateful to Dr. Anabela Valente and Maria Celeste Azevedo of the Universidade de Aveiro for technical assistance with the thermogravimetry analysis. We would also like to acknowledge Dr. Vı´tor Alves of the Instituto Superior de Agronomia for his support in the measurement of the mechanical properties of the membranes, and Dr. Isabel Pombo Cardoso for her help in the visualization of membranes by optical microscopy. References [1] V. Smuleac, J. Wu, S. Nemser, S. Majumdar, D. Bhattacharyya, Novel perfluorinated polymer-based pervaporation membranes for the separation of solvent/water mixtures, J. Membr. Sci. 352 (2010) 41–49.

[2] X. Qiao, T.S. Chung, Fundamental characteristics of sorption, swelling, and permeation of P84 co-polyimide membranes for pervaporation dehydration of alcohols, Ind. Eng. Chem. Res. 44 (2005) 8938–8943. [3] P.D. Chapman, T. Oliveira, A.G. Livingston, K. Li, Membranes for the dehydration of solvents by pervaporation – review, J. Membr. Sci. 318 (2008) 5–37. [4] A. Jonquie res, R. Cle´ment, P. Lochon, J. Ne´el, M. Dresch, B. Chre´tien, Industrial state-of-the-art of pervaporation and vapour permeation in the western countries, J. Membr. Sci. 206 (2002) 87–117. [5] M. Mulder, Basic Principles of Membrane Technology, third ed., Kluwer Academic Publishers, England, 1997. ˜ ez, I. Ortiz, Pervaporation [6] A.M. Urtiaga, E.D. Gorri, P. Go´mez, C. Casado, R. Iba´n technology for the dehydration of solvents and raw materials in the process industry, Drying Technol. 25 (2007) 1819–1828. [7] V. Van Hoof, L. Van den Abeele, A. Buekenhoudt, C. Dotremont, R. Leysen, Economic comparison between azeotropic distillation and different hybrid systems combining distillation with pervaporation for the dehydration of isopropanol, Sep. Purif. Technol. 37 (2004) 33–49. [8] J.G. Varghese, A.A. Kittur, P.S. Rachipudi, M.Y. Kariduraganavar, Synthesis, characterization and pervaporation performance of chitosan-g-polyaniline membranes for the dehydration of isopropanol, J. Membr. Sci. 364 (2010) 111–121. [9] L. Zhang, P. Yu, Y. Luo, Dehydration of caprolactam-water mixtures through cross-linked PVA composite pervaporation membranes, J. Membr. Sci. 306 (2007) 93–102. [10] C. Hu, B. Li, R. Guo, H. Wu, Z. Jiang, Pervaporation performance of chitosanpoly(acrylic acid) polyelectrolyte complex membranes for dehydration of ethylene glycol aqueous solution, Sep. Purif. Technol. 55 (2007) 327–334. [11] Q. Zhao, J. Quian, Q. Na, Z. Zhu, P. Zhang, Y. Bai, Studies on pervaporation characteristics of polyacrylonitrile-b-poly(ethylene glycol)-b-polyacrylonitrile block copolymer membrane for dehydration of aqueous acetone solutions, J. Membr. Sci. 311 (2008) 284–293. [12] Z. Liu, S.Z. Erhan, D.E. Akin, F.E. Barton, Green composites from renewable resources: preparation of epoxidized soybean oil and flax fiber composites, J. Agric. Food Chem. 54 (2006) 2134–2137. [13] L. Yu, K. Dean, L. Li, Polymer blends and composites from renewable resources, Prog. Polym. Sci. 31 (2006) 576–602. [14] K. Zielin´ska, W. Kujawski, A.G. Chostenko, Chitosan hydrogel membranes for pervaporative dehydration of alcohols, Sep. Purif. Technol. 83 (2011) 114–120. [15] S. Kalyani, B. Smitha, S. Sridhar, A. Krishnaiah, Pervaporation separation of ethanol–water mixtures through sodium alginate membranes, Desalination 229 (2008) 68–81. [16] P. Kanti, K. Srigowri, J. Madhuri, B. Smitha, S. Sridhar, Dehydration of ethanol through blend membranes of chitosan and sodium alginate by pervaporation, Sep. Purif. Technol. 40 (2004) 259–266. [17] F. Freitas, V.D. Alves, J. Pais, N. Costa, C. Oliveira, L. Mafra, L. Hilliou, R. Oliveira, M.A.M. Reis, Characterization of an extracellular polysaccharide produced by a Pseudomonas strain grown on glycerol, Bioresour. Technol. 100 (2009) 859–865. [18] F. Freitas, V.D. Alves, M. Carvalheira, N. Costa, R. Oliveira, M.A.M. Reis, Emulsifying behaviour and rheological properties of the extracellular polysaccharide produced by Pseudomonas oleovorans grown on glycerol byproduct, Carbohydr. Polym. 78 (2009) 549–556. [19] V.D. Alves, A.R. Ferreira, N. Costa, F. Freitas, M.A.M. Reis, I.M. Coelhoso, Characterization of biodegradable films from the extracellular polysaccharide produced by Pseudomonas oleovorans grown on glycerol byproduct, Carbohydr. Polym. 83 (2011) 1582–1590. [20] C. Brazinha, J.G. Crespo, Aroma recovery from hydro alcoholic solutions by organophilic pervaporation: modelling of fractionation by condensation, J. Membr. Sci. 341 (2009) 109–121. [21] S. Zereshki, A. Figoli, S.S. Madaeni, S. Simone, J.C. Jansen, M. Esmailinezhad, E. Drioli, Poly(lactic acid)/poly(vinyl pyrrolidone) blend membranes: effect of membrane composition on pervaporation separation of ethanol/cyclohexane mixture, J. Membr. Sci. 362 (2010) 105–112. [22] N.E. Suyatma, A. Copinet, L. Tighzert, V. Coma, Mechanical and barrier properties of biodegradable films made from chitosan and poly(lactic acid) blends, J. Polym. Environ. 12 (1) (2004). [23] J.G. Wijmans, R.W. Baker, The solution-diffusion model: a review, J. Membr. Sci. 107 (1995) 1–21. [24] Site of National Institute of Standards and Technology. Available from: /http://webbook.nist.govS. [25] M.J. Assael, J.P.M. Trusler, T.F. Tsolakis, Thermophysical Properties of Fluids, Imperial College Press (ISBN 1-86094-019-6 (pbk)), 161–206, 1996. [26] R.W. Baker, J.G. Wijmans, Y. Huang, Permeability, permeance and selectivity: a preferred way of reporting pervaporation performance data, J. Membr. Sci. 348 (2010) 346–352. [27] V.V. Namboodiri, L.M. Vane, High permeability membranes for the dehydration of low water content ethanol by pervaporation, J. Membr. Sci. 306 (2007) 209–215. [28] Y.Q. Dong, L. Zhang, J.N. Shen, M.Y. Song, H.L. Chen, Preparation of poly(vinyl alcohol)-sodium alginate hollow-fiber composite membranes and pervaporation dehydration characterization of aqueous alcohol mixtures, Desalination 193 (2006) 202–210.